In an environment in which plural virtual machines are deployed on physical servers and a virtualized system is constructed by causing such plural virtual machines to cooperate each other through a network, it is desired that a test for the virtualized system (hereinafter, also called “virtual system”) is effectively and efficiently carried out. At present, the development and test of such a virtual system are carried out by physically preparing an environment for the development and test. Typically, the scale of the environment for the development and test is smaller than an actually operational environment in which the virtual system will be actually operated, and the environment for the development and test is separated from the actually operational environment.
The physical separation between the actually operational environment and the environment for the development and test allows to exclude the danger of harmfully influencing the actually operational system by the system in the development and test (e.g. the trouble of the system in the development and test causes the stop or mis-operation of the actually operational system). However, there is a case where the actually operational environment is different from the environment for the development and test. For instance, the hardwares of the physical servers may be different, and the software versions of the Operating Systems (OS) or libraries may be different. Therefore, a lot of problems occurs such as troubles, which were not observed in the environment for the development and test, are discovered after the operation of the system in the actually operational environment started, and troubles occur due to the setting mistakes at porting the virtual system from the environment for the development and test to the actually operational environment occur.
Incidentally, various conventional arts exist for the virtual system. However, no technique pays attention to the aforementioned problem.
Recently, non-stop of the business service becomes a very important system requirement. Namely, necessity and importance increase in which the test after the configuration change (e.g. change of the number of virtual machines that cooperate each other, and application of the security patch) of the system is conducted without stopping the business service being actually operated, and the system after the configuration change is promptly made to be actually operated. However, at present, such a requirement is not satisfied. In addition, there is a case where the test for the operational system is desired. However, the dynamic logical configuration change of the virtual system cannot be treated. Furthermore, as for the logical configuration change of the virtual system, various settings are required in the conventional arts, and the load of the service operation manager is large. Therefore, there is possibility that the mis-operation occurs.
Furthermore, although the test typically includes various tests, it is inefficient that all of the tests have to be conducted. Namely, it is preferable that the minimum system test is conducted according to the operation situation of the virtual system.
Accordingly, there is no technique to easily switch the virtual system between the test systems for various types of tests and the operational system in one physical server pool including plural physical servers.